National Grid ESO Deploys ML-based Inertia Forecasting
Great Britain’s electricity network, National Grid ESO, has implemented a machine-learning-based forecasting system from GE Digital giving National Grid ESO a better measure of how much inertia they have in their grid to inform real-time awareness, short-term planning and operational decisions, and long-term investment assessments.
The future of Great Britain’s electricity network looks bright as Electricity System Operator (ESO) National Grid implements a machine-learning-based forecasting system giving National Grid ESO a better measure of how much inertia they have in their grid to inform real-time awareness (is there enough inertia on the grid), short-term planning and operational decisions (will there be enough inertia in June or December) and long-term investment assessments (should a flywheel be procured/built in the northern half of the grid by 2025).
Planning Now for Inertia Challenges
Inertia is a challenge both at the whole-system level and a regional level—and some utilities will have one or both problems, either now or in the near future.
Inertia is what helps keep the grid spinning while backup generation ramps up. GE Digital’s Effective Inertia Metering & Forecasting solution keeps the grid running when a generator suddenly disconnects, making time for other generators to get online. The GE Digital solution will inform National Grid ESO’s engineers of grid inertia day-to-day to help ensure the system can handle a worst-case scenario loss of generators.
The system is built on GE Digital’s energy platform for Wide Area Management (WAMS) applications and is designed to provide system management through enhanced situational awareness, proactive grid management, and maximized transfer capability. The software is developed by GE Digital using industry best practices, considering cybersecurity in design such as incorporating secure APIs, development such as employing code scanning, and operation such as audit logging.
Effective Inertia Metering & Forecasting measures the combined inertia-like effects of rotating machines, passive load responses, and active generator controls through analytics services while WAMS data and analytics measure effective inertia in each regional area of the power system in real-time and can combine them to a system-wide value, according to a press release.
GE Digital’s Effective Inertia Metering & Forecasting provides an operational dashboard that shows operators inertia availability in real-time to minimize risk and increase network resilience. Image used courtesy of GE Digital
Enabling Greener Inertia
Grid inertia is declining with the shift to renewables, since most come from big fossil fuels and nuclear. The leftover grid inertia is expensive and carbon-intensive, as well as harder to measure, with much of it coming from demand and smaller generators.
“If operators don’t have enough inertia in the grid, then the sudden loss of a generator might cause the whole grid to blackout or, rather, a lot of customers disconnected to prevent a complete shutdown of the grid,” said Stuart Clark, Power Systems Engineer, GE Digital, in an interview with EE Power. “If there is too much inertia, then the grid operator has probably procured too much inertia from flywheels and fossil fuel power stations – meaning electricity customers aren’t getting a good deal, and potentially we’ve created a lot of carbon unnecessarily.”
The inertia in the grid changes frequently–hour by hour or even more often–so a good real-time picture is important. The tool will improve National Grid ESO’s ability to manage system stability across the entire network as more renewable energy connects to the grid.
“It…will enable us to generate inertia in a greener way and map and monitor it rather than bringing on coal or gas plants when inertia levels are estimated to be low,” said Julian Leslie, Head of Networks at National Grid ESO in a statement.
The biggest uncertainties existed in the interface between the inertia forecasting system and the various forecasting systems that fed it “predictor” data, such as demand and generation levels.
“GE Digital and National Grid ESO had to work together to develop the interface to share data between these different operational systems–collaborating on the design, then developing from either side in parallel, and finally connecting the systems together,” Clark explained. “We minimized risk on this by collaborating closely on the interface design, and testing as parts of the interface gradually came online (e.g., as the different source forecast systems were connected). This let us spot any problems early and start working on solutions as quickly as possible.”
He continued, “We held regular weekly meetings between the GE Digital and National Grid ESO teams during the project. These meetings enabled us to quickly identify and work to resolve issues…We also felt the benefit of incremental deployment and testing of functionality as it became available instead of testing everything at once near the end of the project.”
Due to the growing volume of variable renewable generation, networks require enhanced system visibility and understanding to deliver fast-acting response services. GE Digital’s Effective Inertia Metering & Forecasting provides confidence to operators to make decisions on appropriate response service. Image used courtesy of GE Digital
The Future of the Grid
This technology is just beginning to be deployed in power grids, as more grids see higher penetrations of inverter-connected energy resources like renewables and HVDC links but more and more utilities are heading beyond the point where they can “install and ignore” inverter-connected generation.
“It is cheaper to think about inertia sooner so you can plan and procure effectively, than have to pay a lot to turn off wind and turn on coal/gas plants to provide inertia when suddenly it becomes a problem on the grid,” Clark said.
The technology can also be used in fast-acting intelligent grid control systems that respond to generator or load disconnections. Knowing the inertia of the region helps these schemes decide how much of a response is needed to stabilize the grid.
GE Digital, National Grid ESO, and the UK National Physical Laboratory are reviewing the performance of the inertia measurements and forecasts which are expected to be completed later in 2022.